Data Cleaning and Analysis

Loading Required Libraries

library(dplyr)
library(tidyr)
library(ggplot2)
library(patchwork)

Reading and Inspecting Data

## old data ##
field_data <- read.csv("~/Remote-sensing/DATA/MF_TreeCensus_Cleaned26082025.csv")
head(field_data)

NA
NA

## new data ##
new_field_data <- read.csv("~/Remote-sensing/DATA/MF_TreeCensus_Cleaned20251002.csv")
head(new_field_data)

Data Validation

OLD Data

#Total amount of data
n_total <- nrow(field_data)

#count distinct tree
#n_tag <- length(unique(field_data$Tag))
n_tag <- n_distinct(field_data$Tag)

#The number of unique (Tag, StemTag) pairs 
n_unique_stem <- nrow(field_data %>% distinct(Tag, StemTag))

n_total

[1] 8568

n_tag

[1] 6505

n_unique_stem

[1] 8535

NEW Data

#Total amount of data
new_n_total <- nrow(new_field_data)

#count distinct tree
#n_tag <- length(unique(field_data$Tag))
new_n_tag <- n_distinct(new_field_data$Tag)

#The number of unique (Tag, StemTag) pairs 
new_unique_stem <- nrow(new_field_data %>% distinct(Tag, StemTag))

new_n_total

[1] 8560

new_n_tag

[1] 6525

new_unique_stem

[1] 8560

## Check for duplicate rows ##
#OLD Data

#Identify rows where (Tag, StemTag) appears more than once
dup_pairs <- field_data %>%
  count(Tag, StemTag, name = "n") %>%
  filter(n > 1)

#View rows with duplicate (Tag, StemTag) pairs 
duplicate_rows <- field_data %>%
  semi_join(dup_pairs, by = c("Tag", "StemTag")) %>%  
  arrange(Tag, StemTag)  
  
duplicate_rows

nrow(duplicate_rows)

[1] 65

## Check for duplicate rows ##
#New Data

#Identify rows where (Tag, StemTag) appears more than once
new_dup_pairs <- new_field_data %>%
  count(Tag, StemTag, name = "n") %>%
  filter(n > 1)

#View rows with duplicate (Tag, StemTag) pairs 
new_duplicate_rows <- new_field_data %>%
  semi_join(new_dup_pairs, by = c("Tag", "StemTag")) %>%  
  arrange(Tag, StemTag)  
  
new_duplicate_rows

nrow(new_duplicate_rows)

[1] 0

#Filter to the first row per group
#OLD Data

df_nodup <- field_data %>%
  group_by(Tag, StemTag) %>%
  slice(1) %>%
  ungroup()

nrow(df_nodup)

[1] 8535

df_nodup

NA

#Filter to the first row per group
#NEW Data
new_df_nodup <- new_field_data %>%
  group_by(Tag, StemTag) %>%
  slice(1) %>%
  ungroup()

nrow(new_df_nodup)

[1] 8560

new_df_nodup

#Check for missing values in each column
#OLD Data
colSums(is.na(df_nodup))

    ExactDate     Plot_name       Quadrat           Tag       StemTag Genus.species            QX            QY            PX 
            0             0             0             0             0             0             0             0             0 
           PY           DBH           HOM          Code 
            0             4            56          6302 

#Check for missing values in each column
#NEW Data
colSums(is.na(new_df_nodup))

    ExactDate     Plot_name       Quadrat           Tag       StemTag Genus_species            QX            QY            PX 
            0             0             0             0             0             0             0             0             0 
           PY           DBH           HOM          Code    Note_Genus 
            0             4            56          1915             0 

#Check rows that have null DBH values
#OLD Data
DBH_na <- df_nodup %>%
  filter(is.na(DBH)) 
head(DBH_na)

NA

#Check rows that have null DBH values
#NEW Data
new_DBH_na <- new_df_nodup %>%
  filter(is.na(DBH)) 


new_stemtag_na <- new_df_nodup %>%
  filter(is.na(StemTag)) 

new_DBH_na

new_stemtag_na

## New Data has no stemtag = 2 rows !!!
# and has no DBH = 4 rows

#CLEAN DATA

#Keep observations with DBH present and DBH ≥ 5
#old data
data_clean <- df_nodup  %>%
  filter(!is.na(DBH) & DBH >= 5)

nrow(data_clean)

[1] 8531

head(data_clean)

#Keep observations with DBH present and DBH ≥ 5 and Stemtag present
#new data

new_data_clean <- new_df_nodup  %>%
  filter(!is.na(DBH) & DBH >= 5 & !is.na(StemTag))

nrow(new_data_clean)

[1] 8556

head(new_data_clean)

VISUALIZATION

#   Identify the species in the dataset 
# OLD DATA #

species_tree <- data_clean %>%
  group_by(Genus.species) %>%
  summarise(count = n())  %>%
  arrange(desc(count))

species_tree

#nrow(species_tree)

#   Identify the species in the dataset 
# New DATA #

new_species_tree <- new_data_clean %>%
  group_by(Genus_species) %>%
  summarise(count = n())  %>%
  arrange(desc(count))

new_species_tree

#nrow(species_tree)

#OLD DATA
# 1) นับจำนวนกิ่งต่อ "ต้น"
stem_count <- data_clean %>%
  group_by(Tag, Genus.species) %>%
  summarise(n_stems = n(), .groups = "drop")

# 2) เลือกเฉพาะต้นที่แตกกิ่ง (มีมากกว่า 1 กิ่ง)
multi_stem <- stem_count %>%
  filter(n_stems > 1)

# 3) นับจำนวน "ต้นที่แตกกิ่ง" ต่อ species
species_multi_count <- multi_stem %>%
  count(Genus.species, sort = TRUE)
  # %>%
  #slice_max(n, n = 12)
species_multi_count

#OLD DATA
species_multi_count <- data_clean %>%
  group_by(Tag, Genus.species) %>%
  summarise(n_stems = n(), .groups = "drop") %>%
  filter(n_stems > 1) %>%
  count(Genus.species, sort = TRUE) 
  #%>%  slice_max(n, n = 12)

species_multi_count

ggplot(species_multi_count, aes(x = reorder(Genus.species, n), y = n)) +
  geom_col(fill = "#26C4B8") +
  coord_flip() +
  labs(
    title = "Number of Multi-stem Trees by Species",
    x = "Species",
    y = "Number of Multi-stem Trees"
  ) +
  theme_minimal(base_size = 12)

# 1) หา Top 12 species (กันซ้ำรายต้นก่อน)
top12_species <- data_clean %>%
  distinct(Tag, Genus.species) %>%
  count(Genus.species, sort = TRUE) %>%
  slice_max(n, n = 12)

# 2) Plot bar chart
ggplot(top12_species, aes(x = reorder(Genus.species, n), y = n)) +
  geom_col(fill = "#26C4B8") +
  coord_flip() +
  labs(
    title = "Top 12 Species by Number of Trees",
    x = "Species",
    y = "Number of Trees"
  ) +
  theme_minimal(base_size = 12)

#NEW DATA
# 1) นับจำนวนกิ่งต่อ "ต้น"
new_stem_count <- new_data_clean %>%
  group_by(Tag, Genus_species) %>%
  summarise(n_stems = n(), .groups = "drop")

# 2) เลือกเฉพาะต้นที่แตกกิ่ง (มีมากกว่า 1 กิ่ง)
new_multi_stem <- new_stem_count %>%
  filter(n_stems > 1)

# 3) นับจำนวน "ต้นที่แตกกิ่ง" ต่อ species
new_species_multi_count <- new_multi_stem %>%
  count(Genus_species, sort = TRUE)

new_species_multi_count

#NEW DATA
new_species_multi_count <- new_data_clean %>%
  group_by(Tag, Genus_species) %>%
  summarise(n_stems = n(), .groups = "drop") %>%
  filter(n_stems > 1) %>%
  count(Genus_species, sort = TRUE)
  # %>% slice_max(n, n = 12)
new_species_multi_count

# ใส่ label ว่าเป็น Old หรือ New
species_multi_count <- species_multi_count %>%
  mutate(Source = "Old")

new_species_multi_count <- new_species_multi_count %>%
  mutate(Source = "New")

# รวมตาราง
combined_multi_count <- bind_rows(
  species_multi_count %>% rename(Species = Genus.species),
  new_species_multi_count %>% rename(Species = Genus_species)
)

library(dplyr)
library(tidyr)
library(ggplot2)
library(stringr)

# 0) ทำความสะอาดชื่อสปีชีส์กันกรณีเว้นวรรค/รูปแบบต่างกัน
combined_multi_count <- combined_multi_count %>%
  mutate(Species = str_squish(Species))

# 1) หา Top 12 จากผลรวม (Old + New)
top12_species <- combined_multi_count %>%
  group_by(Species) %>%
  summarise(Total = sum(n, na.rm = TRUE), .groups = "drop") %>%
  arrange(desc(Total)) %>%
  slice_head(n = 12) %>%
  pull(Species)

# 2) กรองเฉพาะ Top 12 และกำหนดลำดับแกน X ตาม Total (มาก -> น้อย)
species_order <- combined_multi_count %>%
  group_by(Species) %>%
  summarise(Total = sum(n, na.rm = TRUE), .groups = "drop") %>%
  filter(Species %in% top12_species) %>%
  arrange(desc(Total)) %>%
  pull(Species)

combined_top12 <- combined_multi_count %>%
  filter(Species %in% top12_species) %>%
  mutate(Species = factor(Species, levels = species_order)) %>%
  # 3) ให้แต่ละสปีชีส์มีข้อมูล Old และ New อย่างละ 1 แท่ง (ถ้าไม่มี เติม 0)
  complete(Species, Source, fill = list(n = 0))

# กำหนดลำดับ Source = Old ก่อน New
combined_top12 <- combined_top12 %>%
  mutate(Source = factor(Source, levels = c("Old", "New")))

# Plot
ggplot(combined_top12, aes(x = Species, y = n, fill = Source)) +
  geom_col(position = "dodge") +
  labs(
    x = "GenusSpecies",
    y = "Number of multi-stem trees",
    title = "Top 12 species: Old vs New (multi-stem count)"
  ) +
  theme(
    axis.text.x = element_text(angle = 45, hjust = 1),
    legend.position = "top"
  )

ymax <- max(combined_top12$n, na.rm = TRUE)

ggplot(combined_top12, aes(x = Species, y = n, fill = Source)) +
  geom_col(position = "dodge") +
  scale_y_continuous(
    breaks = seq(0, ymax, by = 50),      # กำหนดช่วงทุก 50
    labels = scales::comma               # ใส่ comma เช่น 1,000
  ) +
  labs(
    x = "GenusSpecies",
    y = "Number of multi-stem trees",
    title = "Top 12 species: Old vs New (multi-stem count)"
  ) +
  theme(
    axis.text.x = element_text(angle = 45, hjust = 1),
    legend.position = "top"
  )

#   Bar plot of total tree counts by species
# OLD data

ggplot(species_tree, aes(x = reorder(Genus.species, count), y = count)) +
  geom_col(fill = "steelblue") +
  coord_flip() +  # flip the axis
  labs(title = "Tree count by species ",
       x = "GenusSpecies",
       y = "number of tree") +
  theme_minimal()

#   Bar plot of total tree counts by species
# New data

ggplot(new_species_tree, aes(x = reorder(Genus_species, count), y = count)) +
  geom_col(fill = "steelblue") +
  coord_flip() +  # flip the axis
  labs(title = "Tree count by species ",
       x = "GenusSpecies",
       y = "number of tree") +
  theme_minimal()

#Bar plot of the 12 most abundant species 
#OLD data

top12 <- species_tree %>% slice_head(n = 12)

ggplot(top12, aes(x = reorder(Genus.species, count), y = count)) +
  geom_col(fill = "forestgreen") +
  coord_flip() +
  labs(title = "Top 12 Most common GenusSpecies OLD data ",
       x = "GenusSpecies",
       y = "number of tree") +
  theme_minimal()

#Bar plot of the 12 most abundant species 
#New Data

top12 <- new_species_tree %>% slice_head(n = 12)

ggplot(top12, aes(x = reorder(Genus_species, count), y = count)) +
  geom_col(fill = "forestgreen") +
  coord_flip() +
  labs(title = "Top 12 Most common GenusSpecies New Data",
       x = "GenusSpecies",
       y = "number of tree") +
  theme_minimal()

DBH Distribution by species

DBH for tree level

#Calculate tree-level DBH
# OLD data
dbh_tree <- data_clean %>%
  filter(!is.na(DBH)) %>%             
  group_by(Tag, Genus.species) %>%     
  summarise(DBH_tree = sqrt(sum(DBH^2)), .groups = "drop")

dbh_tree

#Calculate tree-level DBH
# New data
new_dbh_tree <- new_data_clean %>%
  filter(!is.na(DBH)) %>%             
  group_by(Tag, Genus_species) %>%     
  summarise(DBH_tree = sqrt(sum(DBH^2)), .groups = "drop")

new_dbh_tree

# OLD data
# Prepare a list of the top 12 species at the tree level
distribution_species_tree <- dbh_tree %>%
  count(Genus.species, sort = TRUE) %>%
  slice_max(n, n = 12) %>%
  pull(Genus.species)

# Arrange species in ascending order of abundance
species_order_tree <- dbh_tree %>%
  count(Genus.species, sort = TRUE) %>%
  filter(Genus.species %in% distribution_species_tree) %>%
  pull(Genus.species)  %>%
  rev()

# boxplot (tree level)
ggplot(
  dbh_tree %>%
    filter(Genus.species %in% distribution_species_tree) %>%
    mutate(Genus.species = factor(Genus.species, levels = species_order_tree)),
  aes(x = Genus.species, y = DBH_tree)
) +
  geom_boxplot(fill = "#26C4B8", outlier.color = "#E6B800", outlier.size = 1) +
  coord_flip() +
  labs(
    title = "Tree-level DBH Distribution (Top 12 Species) OLD DATA",
    x = "GenusSpecies",
    y = "DBH (cm)"
  ) +
  theme_minimal(base_size = 12)

# New Data
# Prepare a list of the top 12 species at the tree level
distribution_new_species_tree <- new_dbh_tree %>%
  count(Genus_species, sort = TRUE) %>%
  slice_max(n, n = 12) %>%
  pull(Genus_species)

# Arrange species in ascending order of abundance
new_species_order_tree <- new_dbh_tree %>%
  count(Genus_species, sort = TRUE) %>%
  filter(Genus_species %in% distribution_new_species_tree) %>%
  pull(Genus_species)  %>%
  rev()

# boxplot (tree level)
ggplot(
  new_dbh_tree %>%
    filter(Genus_species %in% distribution_new_species_tree) %>%
    mutate(Genus_species = factor(Genus_species, levels = new_species_order_tree)),
  aes(x = Genus_species, y = DBH_tree)
) +
  geom_boxplot(fill = "#E75480", outlier.color = "#E6B800", outlier.size = 1) +
  coord_flip() +
  labs(
    title = "Tree-level DBH Distribution (Top 12 Species) NEW DATA",
    x = "GenusSpecies",
    y = "DBH (cm)"
  ) +
  theme_minimal(base_size = 12)

DBH stem level

OLD data

#The 12 species with the highest number of trees 
distribution_species_stem <- data_clean %>%
  distinct(Tag, Genus.species) %>%         
  count(Genus.species, sort = TRUE) %>% 
  slice_max(n, n = 12) %>%   
  pull(Genus.species)

#   Sequence of the 12 species for plotting, sorted in ascending order of abundance
species_order_stem <- data_clean %>%
  distinct(Tag, Genus.species) %>%
  count(Genus.species, sort = TRUE) %>%
  filter(Genus.species %in% distribution_species_stem) %>%
  pull(Genus.species) %>%
  rev()

# Box plot  for Stem level

ggplot(
  data_clean %>%
    filter(Genus.species %in% distribution_species_stem) %>%
    mutate(Genus.species = factor(Genus.species, levels = species_order_stem)),
  aes(x = Genus.species, y = DBH)
) +
  geom_boxplot(fill = "#26C4B8", outlier.color = "#E6B800", outlier.size = 1) +
  coord_flip() +
  labs(
    title = "Stem-level DBH Distribution (Top 12 Species) OLD DATA",
    x = "GenusSpecies",
    y = "DBH (cm)"
  ) +
  theme_minimal(base_size = 12)

#New data

#The 12 species with the highest number of trees 
new_distribution_species_stem <- new_data_clean %>%
  distinct(Tag, Genus_species) %>%         
  count(Genus_species, sort = TRUE) %>% 
  slice_max(n, n = 12) %>%   
  pull(Genus_species)

#   Sequence of the 12 species for plotting, sorted in ascending order of abundance
new_species_order_stem <- new_data_clean %>%
  distinct(Tag, Genus_species) %>%
  count(Genus_species, sort = TRUE) %>%
  filter(Genus_species %in% new_distribution_species_stem) %>%
  pull(Genus_species) %>%
  rev()

# Box plot  for Stem level

ggplot(
  new_data_clean %>%
    filter(Genus_species %in% new_distribution_species_stem) %>%
    mutate(Genus_species = factor(Genus_species, levels = new_species_order_stem)),
  aes(x = Genus_species, y = DBH)
) +
  geom_boxplot(fill = "#E75480", outlier.color = "#E6B800", outlier.size = 1) +
  coord_flip() +
  labs(
    title = "Stem-level DBH Distribution (Top 12 Species) NEW DATA",
    x = "GenusSpecies",
    y = "DBH (cm)"
  ) +
  theme_minimal(base_size = 12)

#Density plot of DBH by Tree-level
#OLD data
species_order_tree <- dbh_tree %>%
  count(Genus.species, sort = TRUE) %>%
  slice_max(n, n = 12) %>%
  pull(Genus.species)

ggplot(
  dbh_tree %>%
    filter(Genus.species %in% species_order_tree) %>%
    mutate(Genus.species = factor(Genus.species, levels = species_order_tree)),
  aes(x = DBH_tree)
) +
  geom_density(fill = "#26C4B8", alpha = 0.6) +
  facet_wrap(~ Genus.species, scales = "free") + 
  labs(title = "Tree-level DBH Density (Top 12 Species) OLD DATA",
       x = " DBH (cm)", y = "Density") +
  theme_minimal(base_size = 10) +
  theme(legend.position = "none")

#Density plot of DBH by Tree-level
#NEW data
new_species_order_tree <- new_dbh_tree %>%
  count(Genus_species, sort = TRUE) %>%
  slice_max(n, n = 12) %>%
  pull(Genus_species)

ggplot(
  new_dbh_tree %>%
    filter(Genus_species %in% new_species_order_tree) %>%
    mutate(Genus_species = factor(Genus_species, levels = new_species_order_tree)),
  aes(x = DBH_tree)
) +
  geom_density(fill = "#E75480", alpha = 0.6) +
  facet_wrap(~ Genus_species, scales = "free") + 
  labs(title = "Tree-level DBH Density (Top 12 Species) NEW DATA",
       x = " DBH (cm)", y = "Density") +
  theme_minimal(base_size = 10) +
  theme(legend.position = "none")

#Density plot of DBH by Stem-level
#OLD data
species_order_stem <- data_clean %>%
  distinct(Tag, Genus.species) %>%
  count(Genus.species, sort = TRUE) %>%
  filter(Genus.species %in% distribution_species_stem) %>%
  pull(Genus.species)


ggplot(
  data_clean %>%
    filter(Genus.species %in% distribution_species_stem) %>%
    mutate(Genus.species = factor(Genus.species, levels = species_order_stem)),
  aes(x = DBH, fill = Genus.species)
) +
  geom_density(fill = "#26C4B8", alpha = 0.6) +
  facet_wrap(~ Genus.species, scales = "free") +
  labs(
    title = "Stem-level DBH Density (Top 12 Species) OLD DATA",
    x = "DBH (cm)",
    y = "Density"
  ) +
  theme_minimal(base_size = 10) +
  theme(legend.position = "none")

#Density plot of DBH by Stem-level
#NEW data
new_species_order_tree <- new_data_clean %>%
  distinct(Tag, Genus_species) %>%
  count(Genus_species, sort = TRUE) %>%
  filter(Genus_species %in% distribution_new_species_tree) %>%
  pull(Genus_species)


ggplot(
  new_data_clean %>%
    filter(Genus_species %in% distribution_new_species_tree) %>%
    mutate(Genus_species = factor(Genus_species, levels = new_species_order_tree)),
  aes(x = DBH, fill = Genus_species)
) +
  geom_density(fill = "#E75480", alpha = 0.6) +
  facet_wrap(~ Genus_species, scales = "free") +
  labs(
    title = "Stem-level DBH Density (Top 12 Species) NEW DATA",
    x = "DBH (cm)",
    y = "Density"
  ) +
  theme_minimal(base_size = 10) +
  theme(legend.position = "none")